nassp masters 5003f - computational astronomy - 2009 lecture 12 complex numbers – an alternate...

30
NASSP Masters 5003F - Computational Astronomy - 2009 Lecture 12 • Complex numbers – an alternate view • The Fourier transform • Convolution, correlation, filtering.

Upload: cornelius-lamb

Post on 18-Jan-2016

219 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: NASSP Masters 5003F - Computational Astronomy - 2009 Lecture 12 Complex numbers – an alternate view The Fourier transform Convolution, correlation, filtering

NASSP Masters 5003F - Computational Astronomy - 2009

Lecture 12

• Complex numbers – an alternate view

• The Fourier transform

• Convolution, correlation, filtering.

Page 2: NASSP Masters 5003F - Computational Astronomy - 2009 Lecture 12 Complex numbers – an alternate view The Fourier transform Convolution, correlation, filtering

NASSP Masters 5003F - Computational Astronomy - 2009

Complex numbers

REAL IMAGINARY

1iiIRz

Page 3: NASSP Masters 5003F - Computational Astronomy - 2009 Lecture 12 Complex numbers – an alternate view The Fourier transform Convolution, correlation, filtering

NASSP Masters 5003F - Computational Astronomy - 2009

Complex numbers

REAL IMAGINARY

1iiIRz

NONSENSE!

There IS no √-1.

Page 4: NASSP Masters 5003F - Computational Astronomy - 2009 Lecture 12 Complex numbers – an alternate view The Fourier transform Convolution, correlation, filtering

NASSP Masters 5003F - Computational Astronomy - 2009

Let’s ‘forget’ about complex numbers for a bit...

...and talk about 2-component vectors instead.

x

y

v

x

y

sin

cosv

y

xv

θ

Page 5: NASSP Masters 5003F - Computational Astronomy - 2009 Lecture 12 Complex numbers – an alternate view The Fourier transform Convolution, correlation, filtering

NASSP Masters 5003F - Computational Astronomy - 2009

What can we do if we have two of them?

x

y

v1

21

2121 yy

xxvv

v2

We could define something like addition:

There are lots of operations one could define, but only a few of themturn out to be interesting.

vsum

I use a funny symbol to remind us that this is NOTaddition (which is an operationon scalars); it is just analogous to it.:

Page 6: NASSP Masters 5003F - Computational Astronomy - 2009 Lecture 12 Complex numbers – an alternate view The Fourier transform Convolution, correlation, filtering

NASSP Masters 5003F - Computational Astronomy - 2009

The following operation has interesting properties:

x

y

v1

1221

212121prod yxyx

yyxxvvv

v2

But it isn’t very like scalarmultiplication except whenall ys are zero.

21

2121prod sin

cos

vvv

It’s fairly easy to show that:

vprodθ2

θ1

θprod=θ1+θ2

Page 7: NASSP Masters 5003F - Computational Astronomy - 2009 Lecture 12 Complex numbers – an alternate view The Fourier transform Convolution, correlation, filtering

NASSP Masters 5003F - Computational Astronomy - 2009

Vectors? These are just complex numbers!

vv

1

0

Note that:

This, plus the angle-summing properties of theproduct, leads to the following typographicalshorthand:

iv expv

Instead of the mysterious

1iwe should just note the simple identity .

0

1

1

0

1

0

Page 8: NASSP Masters 5003F - Computational Astronomy - 2009 Lecture 12 Complex numbers – an alternate view The Fourier transform Convolution, correlation, filtering

Notation:

NASSP Masters 5003F - Computational Astronomy - 2009

iIRz

izz exp

I

Rz

RIarctan

sincos izz where

These are all just different ways of saying the same thing.

Page 9: NASSP Masters 5003F - Computational Astronomy - 2009 Lecture 12 Complex numbers – an alternate view The Fourier transform Convolution, correlation, filtering

Some important reals:• Phase

• Power

• Amplitude, magnitude or intensity

NASSP Masters 5003F - Computational Astronomy - 2009

RIarctan =atan2(I,R)

22 IRzzzzP

PzA

Page 10: NASSP Masters 5003F - Computational Astronomy - 2009 Lecture 12 Complex numbers – an alternate view The Fourier transform Convolution, correlation, filtering

NASSP Masters 5003F - Computational Astronomy - 2009

The lessons to learn:

• Complex numbers are just 2-vectors.• The ‘imaginary’ part is just as real as the

‘real’ part.• Don’t be fooled by the fact that the same

symbols ‘+’ and ‘x’ are used both for scalar addition/multiplication and for what turn out to be vector operations. This is a historical typographical laziness.– Be aware however that the notation I have

used here, although (IMO) more sensible, is not standard.

– So better go with the flow until you get to be a big shot, and stick with the silly x+iy notation.

Page 11: NASSP Masters 5003F - Computational Astronomy - 2009 Lecture 12 Complex numbers – an alternate view The Fourier transform Convolution, correlation, filtering

The Fourier transform• Analyses a signal into sine and cosines:

• The result is called the spectrum of the signal.NASSP Masters 5003F - Computational Astronomy - 2009

Page 12: NASSP Masters 5003F - Computational Astronomy - 2009 Lecture 12 Complex numbers – an alternate view The Fourier transform Convolution, correlation, filtering

The Fourier transform

• G in general is complex-valued.• ω is an angular frequency (units: radians per unit t).

• the transform is almost self-inverse:

• But remember, these integrals are not guaranteed to converge. (This is not a problem when we ‘compute’ the FT, as will be seen.)

NASSP Masters 5003F - Computational Astronomy - 2009

titgdtGg expF

tiGdtgG exp1-F

Page 13: NASSP Masters 5003F - Computational Astronomy - 2009 Lecture 12 Complex numbers – an alternate view The Fourier transform Convolution, correlation, filtering

Typical transform pairs

NASSP Masters 5003F - Computational Astronomy - 2009

point (delta function) fringes.

By the way, ‘the’ reference for the Fourier transform is Bracewell R, “TheFourier Transform and its Applications”, McGraw-Hill

Page 14: NASSP Masters 5003F - Computational Astronomy - 2009 Lecture 12 Complex numbers – an alternate view The Fourier transform Convolution, correlation, filtering

Typical transform pairs

NASSP Masters 5003F - Computational Astronomy - 2009

‘top hat’ sinc function

Page 15: NASSP Masters 5003F - Computational Astronomy - 2009 Lecture 12 Complex numbers – an alternate view The Fourier transform Convolution, correlation, filtering

Typical transform pairs

NASSP Masters 5003F - Computational Astronomy - 2009

wider narrower

Page 16: NASSP Masters 5003F - Computational Astronomy - 2009 Lecture 12 Complex numbers – an alternate view The Fourier transform Convolution, correlation, filtering

Typical transform pairs

NASSP Masters 5003F - Computational Astronomy - 2009

gaussian gaussian

Page 17: NASSP Masters 5003F - Computational Astronomy - 2009 Lecture 12 Complex numbers – an alternate view The Fourier transform Convolution, correlation, filtering

Typical transform pairs

NASSP Masters 5003F - Computational Astronomy - 2009

Hermitian real

Page 18: NASSP Masters 5003F - Computational Astronomy - 2009 Lecture 12 Complex numbers – an alternate view The Fourier transform Convolution, correlation, filtering

Practical use of the FT:• Periodic signals hidden in noise

• Processing of pure noise:– Correlation– Convolution– Filtering

• Interferometry

NASSP Masters 5003F - Computational Astronomy - 2009

Page 19: NASSP Masters 5003F - Computational Astronomy - 2009 Lecture 12 Complex numbers – an alternate view The Fourier transform Convolution, correlation, filtering

Periodic signal hidden in noise

NASSP Masters 5003F - Computational Astronomy - 2009

The eye can’t see it… …but the transform can.

Page 20: NASSP Masters 5003F - Computational Astronomy - 2009 Lecture 12 Complex numbers – an alternate view The Fourier transform Convolution, correlation, filtering

Transforming pure noise

NASSP Masters 5003F - Computational Astronomy - 2009

Uncorrelated noise The transform looks very similar.This sort of noise is called ‘white’. Why?

Page 21: NASSP Masters 5003F - Computational Astronomy - 2009 Lecture 12 Complex numbers – an alternate view The Fourier transform Convolution, correlation, filtering

Power spectrum• Remember the power P of a complex

number z was defined as

• If we apply this to every complex value of a Fourier spectrum, we get the power spectrum or power spectral density.

• This is both real-valued and positive.• Just as white light contains the same amount

of all frequencies, so does white noise.• (For real data, you have to approximate the

PS by averaging.)NASSP Masters 5003F - Computational Astronomy - 2009

22* IRzzP

Page 22: NASSP Masters 5003F - Computational Astronomy - 2009 Lecture 12 Complex numbers – an alternate view The Fourier transform Convolution, correlation, filtering

Red, brown or 1/f noise

NASSP Masters 5003F - Computational Astronomy - 2009

It’s fractal – looks the sameat all length scales.

Page 23: NASSP Masters 5003F - Computational Astronomy - 2009 Lecture 12 Complex numbers – an alternate view The Fourier transform Convolution, correlation, filtering

Nature…?

NASSP Masters 5003F - Computational Astronomy - 2009

No, it is simulated – 1/f2 noise.

Page 24: NASSP Masters 5003F - Computational Astronomy - 2009 Lecture 12 Complex numbers – an alternate view The Fourier transform Convolution, correlation, filtering

Fourier filtering of noise• Multiply a white spectrum by some band

pass:

• Back-transform:

• The noise is no longer uncorrelated. Now it is correlated noise: ie if the value in one sample is high, this increases the probability that the next sample will also be high.

• I simulated the brown noise in the previous slides via Fourier filtering.

NASSP Masters 5003F - Computational Astronomy - 2009

Page 25: NASSP Masters 5003F - Computational Astronomy - 2009 Lecture 12 Complex numbers – an alternate view The Fourier transform Convolution, correlation, filtering

Another example – bandpass filtering:

NASSP Masters 5003F - Computational Astronomy - 2009

Page 26: NASSP Masters 5003F - Computational Astronomy - 2009 Lecture 12 Complex numbers – an alternate view The Fourier transform Convolution, correlation, filtering

Convolution

NASSP Masters 5003F - Computational Astronomy - 2009

ttgtftdgfth

* =

• It is sort of a smearing/smoothing action.

Page 27: NASSP Masters 5003F - Computational Astronomy - 2009 Lecture 12 Complex numbers – an alternate view The Fourier transform Convolution, correlation, filtering

A very important result:

• This is often a quick way to do a convolution.

• An example of a convolution met already:– Sliding-window linear filters used in source

detection.

NASSP Masters 5003F - Computational Astronomy - 2009

gfgf FFF

Page 28: NASSP Masters 5003F - Computational Astronomy - 2009 Lecture 12 Complex numbers – an alternate view The Fourier transform Convolution, correlation, filtering

Correlation

• It is related to convolution:

• Auto-correlation is the correlation of a function by itself.

• NOTE! For f=noise, this integral will not converge..NASSP Masters 5003F - Computational Astronomy - 2009

ttgtftdgftR gf ,

gfgf FFF

ttftftdfftR ff ,

Page 29: NASSP Masters 5003F - Computational Astronomy - 2009 Lecture 12 Complex numbers – an alternate view The Fourier transform Convolution, correlation, filtering

How to make the autocorrelation converge for a noise signal?

• First recognize that it is often convenient to normalise by dividing by R(0):

• It can be proved that γ(0)=1 and γ(>0)<1.• For ‘sensible’ fs, the following is true:

• A practical calculation estimates equation (1) via some non-infinite value of T. NASSP Masters 5003F - Computational Astronomy - 2009

2tftd

ttftftdt

2

2

2

2

2lim T

T

T

T

T tftd

ttftftdt (1)

Page 30: NASSP Masters 5003F - Computational Astronomy - 2009 Lecture 12 Complex numbers – an alternate view The Fourier transform Convolution, correlation, filtering

Autocorrelation and power spectrum• From slides 9 and 28, it is easy to show

that the Fourier transform of the autocorrelation of a function is the same as its power spectral density.

• Again, in practice, we normalize the PSD by R(0) and estimate the result over a finite bandwidth.

NASSP Masters 5003F - Computational Astronomy - 2009

22 Ffffff FFFF